# Load Necessary Packages
library(estimatr)
library(ggplot2)

# Load Datasets
setwd("/users/josephphillips/Dropbox/COVID-19/Data/FC Data Repository")
usa <- read.csv("USA.csv",header=T,sep=",")
uk <- read.csv("UK.csv",header=T,sep=",")
canada <- read.csv("Canada 1.csv",header=T,sep=",")
canada2 <- read.csv("Canada 2.csv",header=T,sep=",")

# Figure D1
DF1.m1 <- lm_robust(nontarget_false_noleak~W2only+W3only+both+frequent_church+ideo7+lives_highincidence+crt+knowledge+nonwhite+health_trust,data=usa)
DF1.m2 <- lm_robust(W2_nontarget_false_noleak~W2only+W3only+both+nontarget_false_noleak,data=usa)
DF1.m3 <- lm_robust(W3_nontarget_false_noleak~W2only+W3only+both+nontarget_false_noleak,data=usa)
DF1.m4 <- lm_robust(W4_nontarget_false_noleak~W2only+W3only+both+ideo7+nontarget_false_noleak,data=usa)
DF1.m5 <- lm_robust(nontargeted_false_noleak~W2only+W3only+both+knowledge+nonwhite+trust_healthgov,data=uk)
DF1.m6 <- lm_robust(W2_nontargeted_false_noleak~W2only+W3only+both+nontargeted_false_noleak,data=uk)
DF1.m7 <- lm_robust(W3_nontargeted_false_noleak~W2only+W3only+both+knowledge+trust_healthgov+nontargeted_false_noleak,data=uk)
DF1.m8 <- lm_robust(nontargeted_false_noleak~treat+age55+age65+frequentchurch+ideology1+knowledge+nonwhite+polinterest+trust_healthgov+total_media_trust,data=canada)
DF1.m9 <- lm_robust(nontargeted_false_noleak~treat+age45+age55+age65+male+frequentchurch+ideology1+knowledge+nonwhite+trust_healthgov+total_media_trust,data=canada2)
DF2.m1 <- lm_robust(nontarget_true~W2only+W3only+both+male+frequent_church+pid3+ideo7+knowledge+polinterest+health_trust+media_trust,data=usa)
DF2.m2 <- lm_robust(W2_nontarget_true~W2only+W3only+both+ideo7+health_trust+nontarget_true,data=usa)
DF2.m3 <- lm_robust(W3_nontarget_true~W2only+W3only+both+ideo7+health_trust+nontarget_true,data=usa)
DF2.m4 <- lm_robust(W4_nontarget_true~W2only+W3only+both+pid3+ideo7+health_trust+media_trust+nontarget_true,data=usa)
DF2.m5 <- lm_robust(nontargeted_true~W2only+W3only+both+male+knowledge+polinterest+trust_healthgov+total_media_trust,data=uk)
DF2.m6 <- lm_robust(W2_nontargeted_true~W2only+W3only+both+trust_healthgov+nontargeted_true,data=uk)
DF2.m7 <- lm_robust(W3_nontargeted_true~W2only+W3only+both+trust_healthgov+nontargeted_true,data=uk)
DF2.m8 <- lm_robust(nontargeted_true~treat+age55+male+frequentchurch+quebec+west+knowledge+trust_healthgov,data=canada)
DF2.m9 <- lm_robust(nontargeted_true~treat+age45+age55+male+frequentchurch+quebec+ideology1+knowledge+nonwhite+trust_healthgov+total_media_trust,data=canada2)

W2only_plot_data <- subset(data.frame(wave=c(1,2,3,4,1,2,3,2,2),
                                country=factor(c("USA","USA","USA","USA","GB","GB","GB","CA1","CA2"),levels=c("USA","GB","CA1","CA2")),
nfc_noleak=c(DF1.m1$coefficients[2],DF1.m2$coefficients[2],DF1.m3$coefficients[2],DF1.m4$coefficients[2],DF1.m5$coefficients[2],DF1.m6$coefficients[2],DF1.m7$coefficients[2],DF1.m8$coefficients[2],DF1.m9$coefficients[2]),
nfc_lci_noleak=c(DF1.m1$conf.low[2],DF1.m2$conf.low[2],DF1.m3$conf.low[2],DF1.m4$conf.low[2],DF1.m5$conf.low[2],DF1.m6$conf.low[2],DF1.m7$conf.low[2],DF1.m8$conf.low[2],DF1.m9$conf.low[2]),
nfc_uci_noleak=c(DF1.m1$conf.high[2],DF1.m2$conf.high[2],DF1.m3$conf.high[2],DF1.m4$conf.high[2],DF1.m5$conf.high[2],DF1.m6$conf.high[2],DF1.m7$conf.high[2],DF1.m8$conf.high[2],DF1.m9$conf.high[2]),
ntc=c(DF2.m1$coefficients[2],DF2.m2$coefficients[2],DF2.m3$coefficients[2],DF2.m4$coefficients[2],DF2.m5$coefficients[2],DF2.m6$coefficients[2],DF2.m7$coefficients[2],DF2.m8$coefficients[2],DF2.m9$coefficients[2]),
ntc_lci=c(DF2.m1$conf.low[2],DF2.m2$conf.low[2],DF2.m3$conf.low[2],DF2.m4$conf.low[2],DF2.m5$conf.low[2],DF2.m6$conf.low[2],DF2.m7$conf.low[2],DF2.m8$conf.low[2],DF2.m9$conf.low[2]),
ntc_uci=c(DF2.m1$conf.high[2],DF2.m2$conf.high[2],DF2.m3$conf.high[2],DF2.m4$conf.high[2],DF2.m5$conf.high[2],DF2.m6$conf.high[2],DF2.m7$conf.high[2],DF2.m8$conf.high[2],DF2.m9$conf.high[2])),country=="USA" | country=="GB")

canada_plot_data <- subset(data.frame(wave=c(1,2,3,4,1,2,3,2,2),
                               country=factor(c("USA","USA","USA","USA","GB","GB","GB","CA1","CA2"),levels=c("USA","GB","CA1","CA2")),
nfc_noleak=c(DF1.m1$coefficients[2],DF1.m2$coefficients[2],DF1.m3$coefficients[2],DF1.m4$coefficients[2],DF1.m5$coefficients[2],DF1.m6$coefficients[2],DF1.m7$coefficients[2],DF1.m8$coefficients[2],DF1.m9$coefficients[2]),
nfc_lci_noleak=c(DF1.m1$conf.low[2],DF1.m2$conf.low[2],DF1.m3$conf.low[2],DF1.m4$conf.low[2],DF1.m5$conf.low[2],DF1.m6$conf.low[2],DF1.m7$conf.low[2],DF1.m8$conf.low[2],DF1.m9$conf.low[2]),
nfc_uci_noleak=c(DF1.m1$conf.high[2],DF1.m2$conf.high[2],DF1.m3$conf.high[2],DF1.m4$conf.high[2],DF1.m5$conf.high[2],DF1.m6$conf.high[2],DF1.m7$conf.high[2],DF1.m8$conf.high[2],DF1.m9$conf.high[2]),
ntc=c(DF2.m1$coefficients[2],DF2.m2$coefficients[2],DF2.m3$coefficients[2],DF2.m4$coefficients[2],DF2.m5$coefficients[2],DF2.m6$coefficients[2],DF2.m7$coefficients[2],DF2.m8$coefficients[2],DF2.m9$coefficients[2]),
ntc_lci=c(DF2.m1$conf.low[2],DF2.m2$conf.low[2],DF2.m3$conf.low[2],DF2.m4$conf.low[2],DF2.m5$conf.low[2],DF2.m6$conf.low[2],DF2.m7$conf.low[2],DF2.m8$conf.low[2],DF2.m9$conf.low[2]),
ntc_uci=c(DF2.m1$conf.high[2],DF2.m2$conf.high[2],DF2.m3$conf.high[2],DF2.m4$conf.high[2],DF2.m5$conf.high[2],DF2.m6$conf.high[2],DF2.m7$conf.high[2],DF2.m8$conf.high[2],DF2.m9$conf.high[2])),country=="CA1" | country=="CA2")

W3only_plot_data <- subset(data.frame(wave=c(1,2,3,4,1,2,3,2,2),
                               country=factor(c("USA","USA","USA","USA","GB","GB","GB","CA1","CA2"),levels=c("USA","GB","CA1","CA2")),
nfc_noleak=c(DF1.m1$coefficients[3],DF1.m2$coefficients[3],DF1.m3$coefficients[3],DF1.m4$coefficients[3],DF1.m5$coefficients[3],DF1.m6$coefficients[3],DF1.m7$coefficients[3],DF1.m8$coefficients[3],DF1.m9$coefficients[3]),
nfc_lci_noleak=c(DF1.m1$conf.low[3],DF1.m2$conf.low[3],DF1.m3$conf.low[3],DF1.m4$conf.low[3],DF1.m5$conf.low[3],DF1.m6$conf.low[3],DF1.m7$conf.low[3],DF1.m8$conf.low[3],DF1.m9$conf.low[3]),
nfc_uci_noleak=c(DF1.m1$conf.high[3],DF1.m2$conf.high[3],DF1.m3$conf.high[3],DF1.m4$conf.high[3],DF1.m5$conf.high[3],DF1.m6$conf.high[3],DF1.m7$conf.high[3],DF1.m8$conf.high[3],DF1.m9$conf.high[3]),
ntc=c(DF2.m1$coefficients[3],DF2.m2$coefficients[3],DF2.m3$coefficients[3],DF2.m4$coefficients[3],DF2.m5$coefficients[3],DF2.m6$coefficients[3],DF2.m7$coefficients[3],DF2.m8$coefficients[3],DF2.m9$coefficients[3]),
ntc_lci=c(DF2.m1$conf.low[3],DF2.m2$conf.low[3],DF2.m3$conf.low[3],DF2.m4$conf.low[3],DF2.m5$conf.low[3],DF2.m6$conf.low[3],DF2.m7$conf.low[3],DF2.m8$conf.low[3],DF2.m9$conf.low[3]),
ntc_uci=c(DF2.m1$conf.high[3],DF2.m2$conf.high[3],DF2.m3$conf.high[3],DF2.m4$conf.high[3],DF2.m5$conf.high[3],DF2.m6$conf.high[3],DF2.m7$conf.high[3],DF2.m8$conf.high[3],DF2.m9$conf.high[3])),country=="USA" | country=="GB")

both_plot_data <- subset(data.frame(wave=c(1,2,3,4,1,2,3,2,2),
                                     country=factor(c("USA","USA","USA","USA","GB","GB","GB","CA1","CA2"),levels=c("USA","GB","CA1","CA2")),
nfc_noleak=c(DF1.m1$coefficients[4],DF1.m2$coefficients[4],DF1.m3$coefficients[4],DF1.m4$coefficients[4],DF1.m5$coefficients[4],DF1.m6$coefficients[4],DF1.m7$coefficients[4],DF1.m8$coefficients[4],DF1.m9$coefficients[4]),
nfc_lci_noleak=c(DF1.m1$conf.low[4],DF1.m2$conf.low[4],DF1.m3$conf.low[4],DF1.m4$conf.low[4],DF1.m5$conf.low[4],DF1.m6$conf.low[4],DF1.m7$conf.low[4],DF1.m8$conf.low[4],DF1.m9$conf.low[4]),
nfc_uci_noleak=c(DF1.m1$conf.high[4],DF1.m2$conf.high[4],DF1.m3$conf.high[4],DF1.m4$conf.high[4],DF1.m5$conf.high[4],DF1.m6$conf.high[4],DF1.m7$conf.high[4],DF1.m8$conf.high[4],DF1.m9$conf.high[4]),
ntc=c(DF2.m1$coefficients[4],DF2.m2$coefficients[4],DF2.m3$coefficients[4],DF2.m4$coefficients[4],DF2.m5$coefficients[4],DF2.m6$coefficients[4],DF2.m7$coefficients[4],DF2.m8$coefficients[4],DF2.m9$coefficients[4]),
ntc_lci=c(DF2.m1$conf.low[4],DF2.m2$conf.low[4],DF2.m3$conf.low[4],DF2.m4$conf.low[4],DF2.m5$conf.low[4],DF2.m6$conf.low[4],DF2.m7$conf.low[4],DF2.m8$conf.low[4],DF2.m9$conf.low[4]),
ntc_uci=c(DF2.m1$conf.high[4],DF2.m2$conf.high[4],DF2.m3$conf.high[4],DF2.m4$conf.high[4],DF2.m5$conf.high[4],DF2.m6$conf.high[4],DF2.m7$conf.high[4],DF2.m8$conf.high[4],DF2.m9$conf.high[4])),country=="USA" | country=="GB")

cbPalette <- c("#000000", "#E69F00", "#56B4E9", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#CC79A7")

plot_w2only_ntc <- ggplot(W2only_plot_data,aes(x=wave,y=ntc,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("USA","GB","CA1","CA2"),labels=c("USA","GB","Canada 1","Canada 2"),values=cbPalette) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=ntc_lci,ymax=ntc_uci),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",breaks=c(1,2,3,4),labels=c("Pre","Treat","Post","Post")) + ylab("Difference from control") + ggtitle("W2 factcheck only") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)
plot_w3only_ntc <- ggplot(W3only_plot_data,aes(x=wave,y=ntc,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("USA","GB"),values=cbPalette) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=ntc_lci,ymax=ntc_uci),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",breaks=c(1,2,3,4),labels=c("Pre","Pre","Treat","Post")) + ylab("") + ggtitle("W3 factcheck only") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)
plot_both_ntc <- ggplot(both_plot_data,aes(x=wave,y=ntc,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("USA","GB"),values=cbPalette) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=ntc_lci,ymax=ntc_uci),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",breaks=c(1,2,3,4),labels=c("Pre","Treat1","Treat2","Post")) + ylab("") + ggtitle("W2/W3 factchecks") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)
plot_canada_ntc <- ggplot(canada_plot_data,aes(x=wave,y=ntc,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("CA1","CA2"),labels=c("Canada 1","Canada 2"),values=c("#56B4E9","#009E73")) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=ntc_lci,ymax=ntc_uci),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",limits=c(1,4),breaks=c(1,2,3,4),labels=c("N/A","Treat","N/A","N/A")) + ylab("") + ggtitle("Single-wave factcheck") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)

p.ntc <- ggarrange(plot_w2only_ntc,plot_w3only_ntc,plot_both_ntc,plot_canada_ntc,nrow=1,common.legend=T,legend="bottom")

# Table D1
D1.m1 <- lm_robust(W2_nontarget_true~factcheck_treatment,data=usa)
D1.m2 <- lm_robust(W2_nontarget_true~factcheck_treatment+ideo7+health_trust+nontarget_true,data=usa)
D1.m3 <- lm_robust(W3_nontarget_true~(factcheck_treatment*W3factcheck_treatment),data=usa)
D1.m4 <- lm_robust(W3_nontarget_true~(factcheck_treatment*W3factcheck_treatment)+ideo7+health_trust+media_trust+nontarget_true,data=usa)
D1.m5 <- lm_robust(W4_nontarget_true~(factcheck_treatment*W3factcheck_treatment),data=usa)
D1.m6 <- lm_robust(W4_nontarget_true~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+knowledge+health_trust+media_trust+nontarget_true,data=usa)

# Table D2
D2.m1 <- lm_robust(W2_nontargeted_true~factcheck_treatment,data=uk)
D2.m2 <- lm_robust(W2_nontargeted_true~factcheck_treatment+trust_healthgov+nontargeted_true,data=uk)
D2.m3 <- lm_robust(W3_nontargeted_true~(factcheck_treatment*W3factcheck_treatment),data=uk)
D2.m4 <- lm_robust(W3_nontargeted_true~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+nontargeted_true,data=uk)

# Table D3
D3.m1 <- lm_robust(nontargeted_true~treat,data=canada)
D3.m2 <- lm_robust(nontargeted_true~treat+age55+male+frequentchurch+quebec+knowledge+trust_healthgov,data=canada)

# Table D4
D4.m1 <- lm_robust(nontargeted_true~treat,data=canada2)
D4.m2 <- lm_robust(nontargeted_true~treat+age45+age55+age65+male+frequentchurch+quebec+ideology1+knowledge+nonwhite+polinterest+trust_healthgov+total_media_trust+hotspot,data=canada2)

# Table D5
D5.m1 <- lm_robust(W2_accuracy_True1~factcheck_treatment+ideo7+knowledge+health_trust+accuracy_True1,data=usa)
D5.m2 <- lm_robust(W2_accuracy_True2~factcheck_treatment+accuracy_True2,data=usa)
D5.m3 <- lm_robust(W2_accuracy_True3~factcheck_treatment+health_trust+accuracy_True3,data=usa)
D5.m4 <- lm_robust(W2_accuracy_True4~factcheck_treatment+ideo7+health_trust+accuracy_True4,data=usa)
D5.m5 <- lm_robust(W2_accuracy_True5~factcheck_treatment+accuracy_True5,data=usa)
D5.m6 <- lm_robust(W2_accuracy_True6~factcheck_treatment+ideo7+health_trust+accuracy_True6,data=usa)
D5.m7 <- lm_robust(W2_accuracy_change3~factcheck_treatment+pid3+ideo7+health_trust+media_trust+accuracy_change3,data=usa)
D5.m8 <- lm_robust(W2_accuracy_change1~factcheck_treatment+ideo7+health_trust+accuracy_change1,data=usa)

# Table D6
D6.m1 <- lm_robust(W2_accuracy_True1~factcheck_treatment,data=usa)
D6.m2 <- lm_robust(W2_accuracy_True2~factcheck_treatment,data=usa)
D6.m3 <- lm_robust(W2_accuracy_True3~factcheck_treatment,data=usa)
D6.m4 <- lm_robust(W2_accuracy_True4~factcheck_treatment,data=usa)
D6.m5 <- lm_robust(W2_accuracy_True5~factcheck_treatment,data=usa)
D6.m6 <- lm_robust(W2_accuracy_True6~factcheck_treatment,data=usa)
D6.m7 <- lm_robust(W2_accuracy_change3~factcheck_treatment,data=usa)
D6.m8 <- lm_robust(W2_accuracy_change1~factcheck_treatment,data=usa)

# Table D7
D7.m1 <- lm_robust(W3_accuracy_True1~(factcheck_treatment*W3factcheck_treatment)+health_trust+accuracy_True1,data=usa)
D7.m2 <- lm_robust(W3_accuracy_True2~(factcheck_treatment*W3factcheck_treatment)+health_trust+accuracy_True2,data=usa)
D7.m3 <- lm_robust(W3_accuracy_True3~(factcheck_treatment*W3factcheck_treatment)+ideo7+health_trust+accuracy_True3,data=usa)
D7.m4 <- lm_robust(W3_accuracy_True4~(factcheck_treatment*W3factcheck_treatment)+ideo7+health_trust+accuracy_True4,data=usa)
D7.m5 <- lm_robust(W3_accuracy_True5~(factcheck_treatment*W3factcheck_treatment)+accuracy_True5,data=usa)
D7.m6 <- lm_robust(W3_accuracy_True6~(factcheck_treatment*W3factcheck_treatment)+ideo7+knowledge+health_trust+accuracy_True6,data=usa)
D7.m7 <- lm_robust(W3_accuracy_change3~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+health_trust+media_trust+accuracy_change3,data=usa)
D7.m8 <- lm_robust(W3_accuracy_change1~(factcheck_treatment*W3factcheck_treatment)+ideo7+knowledge+health_trust+media_trust+accuracy_change1,data=usa)

# Table D8
D8.m1 <- lm_robust(W3_accuracy_True1~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m2 <- lm_robust(W3_accuracy_True2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m3 <- lm_robust(W3_accuracy_True3~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m4 <- lm_robust(W3_accuracy_True4~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m5 <- lm_robust(W3_accuracy_True5~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m6 <- lm_robust(W3_accuracy_True6~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m7 <- lm_robust(W3_accuracy_change3~(factcheck_treatment*W3factcheck_treatment),data=usa)
D8.m8 <- lm_robust(W3_accuracy_change1~(factcheck_treatment*W3factcheck_treatment),data=usa)

# Table D9
D9.m1 <- lm_robust(W4_accuracy_True1~(factcheck_treatment*W3factcheck_treatment)+ideo7+knowledge+health_trust+accuracy_True1,data=usa)
D9.m2 <- lm_robust(W4_accuracy_True2~(factcheck_treatment*W3factcheck_treatment)+health_trust+media_trust+accuracy_True2,data=usa)
D9.m3 <- lm_robust(W4_accuracy_True3~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+health_trust+media_trust+accuracy_True3,data=usa)
D9.m4 <- lm_robust(W4_accuracy_True4~(factcheck_treatment*W3factcheck_treatment)+ideo7+health_trust+accuracy_True4,data=usa)
D9.m5 <- lm_robust(W4_accuracy_True5~(factcheck_treatment*W3factcheck_treatment)+accuracy_True5,data=usa)
D9.m6 <- lm_robust(W4_accuracy_True6~(factcheck_treatment*W3factcheck_treatment)+ideo7+crt+knowledge+health_trust+accuracy_True6,data=usa)
D9.m7 <- lm_robust(W4_accuracy_change3~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+health_trust+media_trust+accuracy_change3,data=usa)
D9.m8 <- lm_robust(W4_accuracy_change1~(factcheck_treatment*W3factcheck_treatment)+ideo7+knowledge+polinterest+health_trust+media_trust+accuracy_change1,data=usa)

# Table D10
D10.m1 <- lm_robust(W4_accuracy_True1~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m2 <- lm_robust(W4_accuracy_True2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m3 <- lm_robust(W4_accuracy_True3~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m4 <- lm_robust(W4_accuracy_True4~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m5 <- lm_robust(W4_accuracy_True5~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m6 <- lm_robust(W4_accuracy_True6~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m7 <- lm_robust(W4_accuracy_change3~(factcheck_treatment*W3factcheck_treatment),data=usa)
D10.m8 <- lm_robust(W4_accuracy_change1~(factcheck_treatment*W3factcheck_treatment),data=usa)

# Table D11
D11.m1 <- lm_robust(W2_accuracy_True1~factcheck_treatment+trust_healthgov+accuracy_True1,data=uk)
D11.m2 <- lm_robust(W2_accuracy_True2~factcheck_treatment+trust_healthgov+accuracy_True2,data=uk)
D11.m3 <- lm_robust(W2_accuracy_True3~factcheck_treatment+trust_healthgov+accuracy_True3,data=uk)
D11.m4 <- lm_robust(W2_accuracy_True4~factcheck_treatment+trust_healthgov+accuracy_True4,data=uk)
D11.m5 <- lm_robust(W2_accuracy_True5~factcheck_treatment+accuracy_True5,data=uk)
D11.m6 <- lm_robust(W2_accuracy_True6~factcheck_treatment+trust_healthgov+accuracy_True6,data=uk)
D11.m7 <- lm_robust(W2_accuracy_True7~factcheck_treatment+ideology1+trust_healthgov+accuracy_True7,data=uk)
D11.m8 <- lm_robust(W2_accuracy_change1~factcheck_treatment+male+ideology1+knowledge+polinterest+trust_healthgov+accuracy_change1,data=uk)
D11.m9 <- lm_robust(W2_accuracy_change3~factcheck_treatment+accuracy_change3,data=uk)

# Table D12
D12.m1 <- lm_robust(W2_accuracy_True1~factcheck_treatment,data=uk)
D12.m2 <- lm_robust(W2_accuracy_True2~factcheck_treatment,data=uk)
D12.m3 <- lm_robust(W2_accuracy_True3~factcheck_treatment,data=uk)
D12.m4 <- lm_robust(W2_accuracy_True4~factcheck_treatment,data=uk)
D12.m5 <- lm_robust(W2_accuracy_True5~factcheck_treatment,data=uk)
D12.m6 <- lm_robust(W2_accuracy_True6~factcheck_treatment,data=uk)
D12.m7 <- lm_robust(W2_accuracy_True7~factcheck_treatment,data=uk)
D12.m8 <- lm_robust(W2_accuracy_change1~factcheck_treatment,data=uk)
D12.m9 <- lm_robust(W2_accuracy_change3~factcheck_treatment,data=uk)

# Table D13
D13.m1 <- lm_robust(W3_accuracy_True1~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+accuracy_True1,data=uk)
D13.m2 <- lm_robust(W3_accuracy_True2~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+accuracy_True2,data=uk)
D13.m3 <- lm_robust(W3_accuracy_True3~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+accuracy_True3,data=uk)
D13.m4 <- lm_robust(W3_accuracy_True4~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+accuracy_True4,data=uk)
D13.m5 <- lm_robust(W3_accuracy_True5~(factcheck_treatment*W3factcheck_treatment)+male+trust_healthgov+accuracy_True5,data=uk)
D13.m6 <- lm_robust(W3_accuracy_True6~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+accuracy_True6,data=uk)
D13.m7 <- lm_robust(W3_accuracy_True7~(factcheck_treatment*W3factcheck_treatment)+trust_healthgov+accuracy_True7,data=uk)
D13.m8 <- lm_robust(W3_accuracy_change1~(factcheck_treatment*W3factcheck_treatment)+male+accuracy_change1,data=uk)
D13.m9 <- lm_robust(W3_accuracy_change3~(factcheck_treatment*W3factcheck_treatment)+knowledge+polinterest+trust_healthgov+total_media_trust+accuracy_change3,data=uk)

# Table D14
D14.m1 <- lm_robust(W3_accuracy_True1~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m2 <- lm_robust(W3_accuracy_True2~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m3 <- lm_robust(W3_accuracy_True3~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m4 <- lm_robust(W3_accuracy_True4~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m5 <- lm_robust(W3_accuracy_True5~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m6 <- lm_robust(W3_accuracy_True6~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m7 <- lm_robust(W3_accuracy_True7~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m8 <- lm_robust(W3_accuracy_change1~(factcheck_treatment*W3factcheck_treatment),data=uk)
D14.m9 <- lm_robust(W3_accuracy_change3~(factcheck_treatment*W3factcheck_treatment),data=uk)

# Table D15
D15.m1 <- lm_robust(accuracy_True1~treat+age35+age55+age65+male+frequentchurch+west+knowledge+trust_healthgov,data=canada)
D15.m2 <- lm_robust(accuracy_True2~treat+age55+age65+male+frequentchurch+quebec+west+knowledge+nonwhite+trust_healthgov,data=canada)
D15.m3 <- lm_robust(accuracy_True3~treat+age65+frequentchurch+ideology1+trust_healthgov,data=canada)
D15.m4 <- lm_robust(accuracy_True4~treat+age65+male+frequentchurch+quebec+ideology1+knowledge+trust_healthgov,data=canada)
D15.m5 <- lm_robust(accuracy_True5~treat+age65+male+frequentchurch+west+nonwhite+polinterest+trust_healthgov,data=canada)
D15.m6 <- lm_robust(accuracy_True6~treat+quebec+trust_healthgov,data=canada)
D15.m7 <- lm_robust(accuracy_change3~treat+age65+male+frequentchurch+ideology1+trust_healthgov+total_media_trust,data=canada)
D15.m8 <- lm_robust(accuracy_change1~treat+age65+male+quebec+knowledge+polinterest+total_media_trust,data=canada)
D15.m9 <- lm_robust(accuracy_True7~treat+university+age65+ideology1+nonwhite+polinterest+total_media_trust,data=canada)

# Table D16
D16.m1 <- lm_robust(accuracy_True1~treat,data=canada)
D16.m2 <- lm_robust(accuracy_True2~treat,data=canada)
D16.m3 <- lm_robust(accuracy_True3~treat,data=canada)
D16.m4 <- lm_robust(accuracy_True4~treat,data=canada)
D16.m5 <- lm_robust(accuracy_True5~treat,data=canada)
D16.m6 <- lm_robust(accuracy_True6~treat,data=canada)
D16.m7 <- lm_robust(accuracy_change3~treat,data=canada)
D16.m8 <- lm_robust(accuracy_change1~treat,data=canada)
D16.m9 <- lm_robust(accuracy_True7~treat,data=canada)

# Table D17
D17.m1 <- lm_robust(accuracy_True1~treat+age45+age55+male+frequentchurch+ideology1+knowledge+nonwhite+trust_healthgov,data=canada2)
D17.m2 <- lm_robust(accuracy_True2~treat+age45+age55+age65+male+frequentchurch+ideology1+knowledge+nonwhite+trust_healthgov,data=canada2)
D17.m3 <- lm_robust(accuracy_True3~treat+age55+age65+male+frequentchurch+ideology1+knowledge+trust_healthgov,data=canada2)
D17.m4 <- lm_robust(accuracy_True4~treat+male+frequentchurch+quebec+ideology1+knowledge+trust_healthgov,data=canada2)
D17.m5 <- lm_robust(accuracy_True5~treat+male+quebec+ideology1+knowledge+nonwhite+trust_healthgov,data=canada2)
D17.m6 <- lm_robust(accuracy_True6~treat+male+quebec+ideology1+knowledge+nonwhite+polinterest+trust_healthgov+total_media_trust,data=canada2)
D17.m7 <- lm_robust(accuracy_change3~treat+age55+age65+ideology1+trust_healthgov,data=canada2)
D17.m8 <- lm_robust(accuracy_change1~treat,data=canada2)
D17.m9 <- lm_robust(accuracy_True7~treat+university+left+polinterest+trust_healthgov+total_media_trust,data=canada2)

# Table D18
D18.m1 <- lm_robust(accuracy_True1~treat,data=canada2)
D18.m2 <- lm_robust(accuracy_True2~treat,data=canada2)
D18.m3 <- lm_robust(accuracy_True3~treat,data=canada2)
D18.m4 <- lm_robust(accuracy_True4~treat,data=canada2)
D18.m5 <- lm_robust(accuracy_True5~treat,data=canada2)
D18.m6 <- lm_robust(accuracy_True6~treat,data=canada2)
D18.m7 <- lm_robust(accuracy_change3~treat,data=canada2)
D18.m8 <- lm_robust(accuracy_change1~treat,data=canada2)
D18.m9 <- lm_robust(accuracy_True7~treat,data=canada2)

# Figure D2 (see Figure D1 code for models)
plot_w2only_nfc_noleak <- ggplot(W2only_plot_data,aes(x=wave,y=nfc_noleak,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("USA","GB","CA1","CA2"),labels=c("USA","GB","Canada 1","Canada 2"),values=cbPalette) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=nfc_lci_noleak,ymax=nfc_uci_noleak),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",breaks=c(1,2,3,4),labels=c("Pre","Treat","Post","Post")) +  ylab("Difference from control") + ggtitle("W2 factcheck only") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)
plot_w3only_nfc_noleak <- ggplot(W3only_plot_data,aes(x=wave,y=nfc_noleak,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("USA","GB"),values=cbPalette) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=nfc_lci_noleak,ymax=nfc_uci_noleak),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",breaks=c(1,2,3,4),labels=c("Pre","Pre","Treat","Post")) + ylab("") + ggtitle("W3 factcheck only") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)
plot_both_nfc_noleak <- ggplot(both_plot_data,aes(x=wave,y=nfc_noleak,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("USA","GB"),values=cbPalette) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=nfc_lci_noleak,ymax=nfc_uci_noleak),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",breaks=c(1,2,3,4),labels=c("Pre","Treat1","Treat2","Post")) + ylab("") + ggtitle("W2/W3 factchecks") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)
plot_canada_nfc_noleak <- ggplot(canada_plot_data,aes(x=wave,y=nfc_noleak,group=country,color=country)) + geom_point(position=position_dodge(0.2)) + scale_color_manual(name="Country",limits=c("CA1","CA2"),labels=c("Canada 1","Canada 2"),values=c("#56B4E9","#009E73")) + geom_line(linetype="dashed",position=position_dodge(0.2)) + geom_errorbar(aes(ymin=nfc_lci_noleak,ymax=nfc_uci_noleak),size=.4,width=0,position=position_dodge(0.2)) + theme_classic() + geom_hline(yintercept=0,linetype="dashed",color="gray40") + scale_x_continuous(name="Wave",limits=c(1,4),breaks=c(1,2,3,4),labels=c("N/A","Treat","N/A","N/A")) + ylab("") + ggtitle("Single-wave factcheck") + theme(plot.title = element_text(hjust = 0.5)) + ylim(-0.48,0.28)

p.nfc <- ggarrange(plot_w2only_nfc_noleak,plot_w3only_nfc_noleak,plot_both_nfc_noleak,plot_canada_nfc_noleak,nrow=1,common.legend=T,legend="bottom")

# Table D19
D19.m1 <- lm_robust(W2_nontarget_false_noleak~factcheck_treatment,data=usa)
D19.m2 <- lm_robust(W2_nontarget_false_noleak~factcheck_treatment+nontarget_false_noleak,data=usa)
D19.m3 <- lm_robust(W3_nontarget_false_noleak~factcheck_treatment,data=usa)
D19.m4 <- lm_robust(W3_nontarget_false_noleak~(factcheck_treatment*W3factcheck_treatment)+nontarget_false_noleak,data=usa)
D19.m5 <- lm_robust(W4_nontarget_false_noleak~factcheck_treatment,data=usa)
D19.m6 <- lm_robust(W4_nontarget_false_noleak~(factcheck_treatment*W3factcheck_treatment)+ideo7+nontarget_false_noleak,data=usa)

# Table D20
D20.m1 <- lm_robust(W2_nontargeted_false_noleak~factcheck_treatment,data=uk)
D20.m2 <- lm_robust(W2_nontargeted_false_noleak~factcheck_treatment+nontargeted_false_noleak,data=uk)
D20.m3 <- lm_robust(W3_nontargeted_false_noleak~(factcheck_treatment*W3factcheck_treatment),data=uk)
D20.m4 <- lm_robust(W3_nontargeted_false_noleak~(factcheck_treatment*W3factcheck_treatment)+knowledge+trust_healthgov+nontargeted_false_noleak,data=uk)

# Table D21
D21.m1 <- lm_robust(nontargeted_false_noleak~treat,data=canada)
D21.m2 <- lm_robust(nontargeted_false_noleak~treat+age55+age65+frequentchurch+ideology1+knowledge+nonwhite+polinterest+trust_healthgov+total_media_trust,data=canada)

# Table D22
D22.m1 <- lm_robust(nontargeted_false_noleak~treat,data=canada2)
D22.m2 <- lm_robust(nontargeted_false_noleak~treat+age45+age55+age65+male+frequentchurch+left+ideology1+knowledge+nonwhite,data=canada2)

# Table D23
D23.m1 <- lm_robust(W2_accuracy_False2~factcheck_treatment+accuracy_False2,data=usa)
D23.m2 <- lm_robust(W2_accuracy_False3~factcheck_treatment+accuracy_False3,data=usa)
D23.m3 <- lm_robust(W2_accuracy_edge1~factcheck_treatment+knowledge+accuracy_edge1,data=usa)
D23.m4 <- lm_robust(W2_accuracy_f_cure1~factcheck_treatment+accuracy_f_cure1,data=usa)
D23.m5 <- lm_robust(W2_accuracy_f_cure2~factcheck_treatment+ideo7+crt+health_trust+accuracy_f_cure2,data=usa)
D23.m6 <- lm_robust(W2_accuracy_f_cure3~factcheck_treatment+ideo7+accuracy_f_cure3,data=usa)

# Table D24
D24.m1 <- lm_robust(W2_accuracy_False2~factcheck_treatment,data=usa)
D24.m2 <- lm_robust(W2_accuracy_False3~factcheck_treatment,data=usa)
D24.m3 <- lm_robust(W2_accuracy_edge1~factcheck_treatment,data=usa)
D24.m4 <- lm_robust(W2_accuracy_f_cure1~factcheck_treatment,data=usa)
D24.m5 <- lm_robust(W2_accuracy_f_cure2~factcheck_treatment,data=usa)
D24.m6 <- lm_robust(W2_accuracy_f_cure3~factcheck_treatment,data=usa)

# Table D25
D25.m1 <- lm_robust(W3_accuracy_False2~(factcheck_treatment*W3factcheck_treatment)+accuracy_False2,data=usa)
D25.m2 <- lm_robust(W3_accuracy_False3~(factcheck_treatment*W3factcheck_treatment)+accuracy_False3,data=usa)
D25.m3 <- lm_robust(W3_accuracy_edge1~(factcheck_treatment*W3factcheck_treatment)+accuracy_edge1,data=usa)
D25.m4 <- lm_robust(W3_accuracy_f_cure1~(factcheck_treatment*W3factcheck_treatment)+accuracy_f_cure1,data=usa)
D25.m5 <- lm_robust(W3_accuracy_f_cure2~(factcheck_treatment*W3factcheck_treatment)+ideo7+health_trust+accuracy_f_cure2,data=usa)
D25.m6 <- lm_robust(W3_accuracy_f_cure3~(factcheck_treatment*W3factcheck_treatment)+ideo7+knowledge+health_trust+accuracy_f_cure3,data=usa)

# Table D26
D26.m1 <- lm_robust(W3_accuracy_False2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D26.m2 <- lm_robust(W3_accuracy_False3~(factcheck_treatment*W3factcheck_treatment),data=usa)
D26.m3 <- lm_robust(W3_accuracy_edge1~(factcheck_treatment*W3factcheck_treatment),data=usa)
D26.m4 <- lm_robust(W3_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D26.m5 <- lm_robust(W3_accuracy_f_cure1~(factcheck_treatment*W3factcheck_treatment),data=usa)
D26.m6 <- lm_robust(W3_accuracy_f_cure2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D26.m7 <- lm_robust(W3_accuracy_f_cure3~(factcheck_treatment*W3factcheck_treatment),data=usa)

# Table D27
D27.m1 <- lm_robust(W4_accuracy_False2~(factcheck_treatment*W3factcheck_treatment)+accuracy_False2,data=usa)
D27.m2 <- lm_robust(W4_accuracy_False3~(factcheck_treatment*W3factcheck_treatment)+accuracy_False3,data=usa)
D27.m3 <- lm_robust(W4_accuracy_edge1~(factcheck_treatment*W3factcheck_treatment)+accuracy_edge1,data=usa)
D27.m4 <- lm_robust(W4_accuracy_f_cure1~(factcheck_treatment*W3factcheck_treatment)+accuracy_f_cure1,data=usa)
D27.m5 <- lm_robust(W4_accuracy_f_cure2~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+health_trust+accuracy_f_cure2,data=usa)
D27.m6 <- lm_robust(W4_accuracy_f_cure3~(factcheck_treatment*W3factcheck_treatment)+ideo7+knowledge+health_trust+accuracy_f_cure3,data=usa)

# Table D28
D28.m1 <- lm_robust(W4_accuracy_False2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D28.m2 <- lm_robust(W4_accuracy_False3~(factcheck_treatment*W3factcheck_treatment),data=usa)
D28.m3 <- lm_robust(W4_accuracy_edge1~(factcheck_treatment*W3factcheck_treatment),data=usa)
D28.m4 <- lm_robust(W4_accuracy_f_cure1~(factcheck_treatment*W3factcheck_treatment),data=usa)
D28.m5 <- lm_robust(W4_accuracy_f_cure2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D28.m6 <- lm_robust(W4_accuracy_f_cure3~(factcheck_treatment*W3factcheck_treatment),data=usa)

# Table D29
D29.m1 <- lm_robust(W2_accuracy_False2~factcheck_treatment+accuracy_False2,data=uk)
D29.m2 <- lm_robust(W2_accuracy_edge1~factcheck_treatment+accuracy_edge1,data=uk)
D29.m3 <- lm_robust(W2_accuracy_f_cure1~factcheck_treatment+accuracy_f_cure1,data=uk)
D29.m4 <- lm_robust(W2_accuracy_f_cure2~factcheck_treatment+accuracy_f_cure2,data=uk)

# Table D30
D30.m1 <- lm_robust(W2_accuracy_False2~factcheck_treatment,data=uk)
D30.m2 <- lm_robust(W2_accuracy_edge1~factcheck_treatment,data=uk)
D30.m3 <- lm_robust(W2_accuracy_f_cure1~factcheck_treatment,data=uk)
D30.m4 <- lm_robust(W2_accuracy_f_cure2~factcheck_treatment,data=uk)

# Table D31
D31.m1 <- lm_robust(W3_accuracy_False2~(factcheck_treatment*W3factcheck_treatment)+accuracy_False2,data=uk)
D31.m2 <- lm_robust(W3_accuracy_edge1~(factcheck_treatment*W3factcheck_treatment)+accuracy_edge1,data=uk)
D31.m3 <- lm_robust(W3_accuracy_f_cure1~(factcheck_treatment*W3factcheck_treatment)+knowledge+trust_healthgov+accuracy_f_cure1,data=uk)
D31.m4 <- lm_robust(W3_accuracy_f_cure2~(factcheck_treatment*W3factcheck_treatment)+knowledge+trust_healthgov+accuracy_f_cure2,data=uk)

# Table D32
D32.m1 <- lm_robust(W3_accuracy_False2~(factcheck_treatment*W3factcheck_treatment),data=uk)
D32.m2 <- lm_robust(W3_accuracy_edge1~(factcheck_treatment*W3factcheck_treatment),data=uk)
D32.m3 <- lm_robust(W3_accuracy_f_cure1~(factcheck_treatment*W3factcheck_treatment),data=uk)
D32.m4 <- lm_robust(W3_accuracy_f_cure2~(factcheck_treatment*W3factcheck_treatment),data=uk)

# Table D33
D33.m1 <- lm_robust(accuracy_False2~treat,data=canada)
D33.m2 <- lm_robust(accuracy_False3~treat+frequentchurch+quebec+ideology1+knowledge+nonwhite+polinterest,data=canada)
D33.m3 <- lm_robust(accuracy_edge1~treat+age55+age65+male+frequentchurch+quebec+west+ideology1+knowledge+nonwhite+polinterest+trust_healthgov,data=canada)
D33.m4 <- lm_robust(accuracy_f_cure_1~treat+age55+age65+male+frequentchurch+quebec+ideology1+knowledge+nonwhite+polinterest+trust_healthgov,data=canada)
D33.m5 <- lm_robust(accuracy_f_cure_2~treat+age55+age65+ontario+frequentchurch+ideology1+nonwhite+polinterest+trust_healthgov,data=canada)
D33.m6 <- lm_robust(accuracy_f_cure_3~treat+age45+age55+age65+frequentchurch+quebec+ideology1+knowledge+nonwhite+polinterest+trust_healthgov,data=canada)

# Table D34
D34.m1 <- lm_robust(accuracy_False2~treat,data=canada)
D34.m2 <- lm_robust(accuracy_False3~treat,data=canada)
D34.m3 <- lm_robust(accuracy_edge1~treat,data=canada)
D34.m4 <- lm_robust(accuracy_con_theory2~treat,data=canada)
D34.m5 <- lm_robust(accuracy_f_cure_1~treat,data=canada)
D34.m6 <- lm_robust(accuracy_f_cure_2~treat,data=canada)
D34.m7 <- lm_robust(accuracy_f_cure_3~treat,data=canada)

# Table D35
D35.m1 <- lm_robust(accuracy_False2~treat+age65+frequentchurch+quebec+ideology1+nonwhite+total_media_trust,data=canada2)
D35.m2 <- lm_robust(accuracy_False3~treat+age45+age65+male+frequentchurch+quebec+left+ideology1+knowledge+nonwhite,data=canada2)
D35.m3 <- lm_robust(accuracy_edge1~treat+age45+age55+age65+frequentchurch+left+ideology1+knowledge+nonwhite+trust_healthgov,data=canada2)
D35.m4 <- lm_robust(accuracy_con_theory2~treat+ideology1+trust_healthgov,data=canada2)
D35.m5 <- lm_robust(accuracy_f_cure_1~treat+age65+frequentchurch+ideology1+knowledge+nonwhite,data=canada2)
D35.m6 <- lm_robust(accuracy_f_cure_2~treat+age45+age65+frequentchurch+ontario+ideology1+knowledge+nonwhite+trust_healthgov,data=canada2)
D35.m7 <- lm_robust(accuracy_f_cure_3~treat+age45+age55+age65+male+frequentchurch+left+ideology1+knowledge+nonwhite+trust_healthgov,data=canada2)

# Table D36
D36.m1 <- lm_robust(accuracy_False2~treat,data=canada2)
D36.m2 <- lm_robust(accuracy_False3~treat,data=canada2)
D36.m3 <- lm_robust(accuracy_edge1~treat,data=canada2)
D36.m4 <- lm_robust(accuracy_con_theory2~treat,data=canada2)
D36.m5 <- lm_robust(accuracy_f_cure_1~treat,data=canada2)
D36.m6 <- lm_robust(accuracy_f_cure_2~treat,data=canada2)
D36.m7 <- lm_robust(accuracy_f_cure_3~treat,data=canada2)

# Table D37
D37.m1 <- lm_robust(W2_accuracy_con_theory2~factcheck_treatment,data=usa)
D37.m2 <- lm_robust(W2_accuracy_con_theory2~factcheck_treatment+pid3+ideo7+health_trust+accuracy_con_theory2,data=usa)
D37.m3 <- lm_robust(W3_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D37.m4 <- lm_robust(W3_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+health_trust+media_trust+accuracy_con_theory2,data=usa)
D37.m5 <- lm_robust(W4_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment),data=usa)
D37.m6 <- lm_robust(W4_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment)+pid3+ideo7+health_trust+media_trust+accuracy_con_theory2,data=usa)

# Table D38
D38.m1 <- lm_robust(W2_accuracy_con_theory2~factcheck_treatment,data=uk)
D38.m2 <- lm_robust(W2_accuracy_con_theory2~factcheck_treatment+ideology1+trust_healthgov+accuracy_con_theory2,data=uk)
D38.m3 <- lm_robust(W3_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment),data=uk)
D38.m4 <- lm_robust(W3_accuracy_con_theory2~(factcheck_treatment*W3factcheck_treatment)+ideology1+trust_healthgov+accuracy_con_theory2,data=uk)

# Table D39
D39.m1 <- lm_robust(accuracy_con_theory2~treat,data=canada)
D39.m2 <- lm_robust(accuracy_con_theory2~treat+university+age65+frequentchurch+nonwhite+ideology1+polinterest+trust_healthgov+total_media_trust,data=canada)

# Table D40
D40.m1 <- lm_robust(accuracy_con_theory2~treat,data=canada2)
D40.m2 <- lm_robust(accuracy_con_theory2~treat+ideology1+trust_healthgov,data=canada2)